Offline Deep Reinforcement Learning for Dynamic Pricing of Consumer Credit

TL;DR

This paper presents an offline deep reinforcement learning method for personalized consumer credit pricing, leveraging static data and conservative Q-Learning to optimize prices without online testing.

Contribution

It introduces a novel application of offline deep reinforcement learning, specifically conservative Q-Learning, for dynamic credit pricing without requiring online interaction.

Findings

Effective personalized pricing policy learned from static data

No need for online price experimentation

Works on both real and synthetic datasets

Abstract

We introduce a method for pricing consumer credit using recent advances in offline deep reinforcement learning. This approach relies on a static dataset and requires no assumptions on the functional form of demand. Using both real and synthetic data on consumer credit applications, we demonstrate that our approach using the conservative Q-Learning algorithm is capable of learning an effective personalized pricing policy without any online interaction or price experimentation.